sad really but it is overpowered and the player would only forget about the smaller ships.

The problem as I see it in the video are the turrets.

They traverse super-fast and so they have no trouble bearing on "the smaller ships".

If turrets traverse very slowly

• they feel more massive
• they can not hope to track small ships that get close.

-> Vulnerability: Small ships.

If small ships are also fast and evading laterally, slower "capital ship" projectiles can help.

Capital ships wouldn't be agile enough to dance around the bullets but a corvette might be able to at long to medium range... until it's too close for the BB turrets to keep tracking it.

This would lead to capital ships requiring a screen of destroyers or something similarly sized.

That's not to say that a BB must not have any smaller guns at all. It would just have to be a very limited capability like one small turret on bow and stern.

Basically every ship needs a role and... even more important... clearly defined vulnerabilities.

Its terribly easy to build perfect ships - not least because it makes sense to do so - but they make for very boring gameplay. =)

Now, now traversing speed isn't an automatic fix because it demands a smarter AI.

Otherwise you'll see what happens in older versions of X3:

A big and slow turret locks on to something fast that it cannot realistically hit.

It does not reconsider that decision but keeps tracking that tiny drone or whatever... even when a hostile battleship has since come in range.

The latest X3 (AP) runs my turret script and that works differently. The player can assign his own list of priorities for what the turret should attack and in which order. If fighter drones are not in the list, it's just not interested no matter how hard the drone tries to draw fire.

These priorities are re-evaluated once in a while (even while a target is active) so a turret doesn't "lock up" on a sub-optimal target.

Mind you, I'm not asking to be able to assign turret target priorities in SS. Not for a RTS game.

Just saying that there is no such thing as an easy fix when you're dealing with game balance. =)

I understand your points but I'm balancing playability with realism and I'm afraid there will be casualties. Everything you see will be tweaked as the game evolves but it would take years for a team of programmers to create the behaviour you describe. Think of Star Citizen and the level of complexity they are trying to fold in; they have a budget of millions, a strong team of developers and still it's likely they are having difficulties else we'd see something playable by now. I am one guy at the minute so I'm frantically working to get a prototype I can use to raise cash to work on this full-time.

Ships moving around evading fire? Then i'll need to write complex collision avoidance mechanisms munching your CPU.

Slow turrets tracking down? Then the battles will take 3-4x the time to play out.

I haven't just blindly draw the conclusions you've seen on screen, I've been testing and adjusting for 6 months now - I joke that this game has a soul and due to the 3 dimensions of movement the style of play is inferred to a certain degree else it will be confusing. That said I've sank hundreds of hours into X3:TC and X3:AP myself so I know what I would like to see also, but don't forget that X3 was unplayable for the majority and that is not something I care to emulate.

The player will be able to configure the ships for different roles, the turrets will be configurable in a 4X fashion (as will other ship modules.) There are 4 types of turret (Small, Medium, Large, Very Large) designed to combat the various types of ships (Corvette/Fighter, Frigate, Cruiser, Capital Cruiser) the turrets themselves know which ships to prioritise to take that complexity away from the player with the larger turrets unable to fire at all on the smaller ships. The AI is programmed to use the turrets effectively even now, so you won't see a AI frigate equipped as an anti-fighter platform chasing cruisers (unless there are no fighters in play.)

Small turrets deal good damage to Small targets, a reasonable amount of damage to Medium targets and next to no damage to large targets so it will be important to use the right ships for the right job. With that in mind the player will be forced to use smaller ships to escort the larger ships and will need to design ships suitable for certain tasks.

That's pretty much what I was asking for - only worded differently. =)
 
 

Ships moving around evading fire? Then i'll need to write complex collision avoidance mechanisms munching your CPU.

IMO that's mostly a matter of scale.
If the opponents are impossibly close to one another, then yes, collisions will play a role.

With decent bullet range and speed, fights could happen at longer ranges and evasion wouldn't be such a drain - or tiny amount of clipping would be less noticeable. =)

Oh fair enough then 🙂 I find sometimes I run away with thoughts in these posts rather than answering the question, I would make a good politician.

The trouble is that the smaller ships do move around and the larger ships move to get the most out of their heaviest turrets, collision avoidance because of this is really complicated to work out. Each ship would need to be aware of other ships trajectories and alter their speed and/or direction to compensate, taking into account the potential colliding ships dimensions; headfuck!

I fear that sort of maths is beyond my skills but maybe I'll come across something or have an epiphany - too much of a time sink at the moment though.

The trouble with fights happening over greater distances is the field of view of the camera, the greater the distances in the engagement the smaller the ships become onscreen due to perspective which dulls the experience somewhat. So I'm having to use the max weapons ranges to lasso the ships together in battle so they look better on-screen.

Very much an evolutionary title!

About collision avoidance.

http://www.eng.auburn.edu/files/acad_depts/csse/csse_technical_reports/csse11-03.pdf

http://movingai.com/mapf/slides/manocha.pdf

http://grail.cs.washington.edu/projects/crowd-flows/

http://www.insaneunity.com/blog/2011/11/09/artificial-intelligence-part-1-path-finding-with-collision-avoidance/

http://buildnewgames.com/vector-field-collision-avoidance/

I think that the flow fields would do the trick.

Would be this helpfull at some extent?

Best regards

Oh my that is very helpful actually, I'm liking the look of Optimal Reciprocal Collision Avoidance and am pleased there is some source code out there to reference.

Which link refers to flow fields? I couldn't see...

Which link refers to flow fields? I couldn't see...

The first PDF, the crowd-flows (the third link), the fourth and the fifth.

Flow Fields, Vector Fields, Magnetic Fields, all these are names for essentially the same thing, when we are talking about collision avoidance.

Flow fields, you know, you assign a vector to each "map point", the speed of each agent is the actual XYZ position + the VxVyVz of the vector at this position in the field. All these vectors are calculated based on the obstacles to be avoided. Basicly, target positions attracts the agents, obstacles, repulse the agents. You would view this as magnetic field too. A destination attracts, an obstacle repulses.

You would need to calculate the field at each game cycle, because your obstacles are the same ships, that they are in movement. Well, you know this way better than me. A trick that you would use to speed up the thing is recalculate this field only for ships in movement, dividing this field into zones.

I don't know, I just try giving you some inspiration, since that I was programmer (ASP / SQL / Business Applications) years ago, but not games programmer, this is a complete different beast.

Best regards.

I think with some tweaking, and liberal use of "bubbles" around ships you could make flow fields work well enough for this. You'd still need to be careful to not make it floaty or lightweight in terms of how ships behave when encountering another. You wouldn't want a larger ship moving at full speed to be able to quickly adjust it's pathing, which would probably leave you with a behavior system similar to something like Eve Online (where both ships end up reacting to the "collision" based on relative size/mass. Fighters could easily slide along a battleship without affecting the battleship, but two battleships colliding would mean both adjusting paths even if both aren't moving.)

This all really depends on how nice you want the reactive pathing to look. You could easily get away with a system very much like SupCom2.

Supreme Commander 2 is a popular example of flow field pathing due to it's scale. This video, while not particularly informative on the programming side, shows what an end product would look like minus adjustments made to fit your particular case.

Flow fields, in my opinion, are primarily useful in instances with large amounts of units or pathfinding taking place simultaneously as it allows the units to... well... flow past each other without rerouting upon detecting an obstacle (another moving unit) suddenly in it's pre-determined path. I suspect it's also easier on the system running it.

*Am not a programmer of any description*

That's incredibly handy SleepyJo! Very useful for when I come round to it, thanks for taking the time to write that out.

This forum is becoming a very useful collection of ideas 🙂 keep 'em coming.